Table 6.

The performance of each deep learning model in the MVI prediction task.

Model	Scan Time Period	Acc	Rec	Prec	F1 Score
MVI-Mind	AP (avg ± 95%CI)	0.8678	0.8269	0.8750	0.8488
		±0.0458	±0.0767	±0.0490	±0.0566
	PP (avg ± 95%CI)	0.8534	0.7760	0.8972	0.8241
		±0.0484	±0.1060	±0.0651	±0.0645
	DP (avg ± 95%CI)	0.8434	0.7637	0.8823	0.8150
		±0.0547	±0.0802	±0.0816	±0.0660
ResNet-34	AP (avg ± 95%CI)	0.8283	0.6988	0.9089	0.7875
		±0.0242	±0.0372	±0.0676	±0.0303
	PP (avg ± 95%CI)	0.7844	0.6684	0.8313	0.7356
		±0.0474	±0.0905	±0.0732	±0.0688
	DP (avg ± 95%CI)	0.7889	0.6848	0.8271	0.7409
		±0.0653	±0.1242	±0.0834	±0.0919
Inception-V3	AP (avg ± 95%CI)	0.7940	0.7256	0.8061	0.7599
		±0.0269	±0.0738	±0.0478	±0.0439
	PP (avg ± 95%CI)	0.7728	0.6949	0.7911	0.7380
		±0.0525	±0.0450	±0.0787	±0.0512
	DP (avg ± 95%CI)	0.7947	0.7423	0.8133	0.7682
		±0.0501	±0.0793	±0.0948	±0.0494

Note: Each model’s results are the mean of 5 predictions and the corresponding 95% confidence interval.